REM WORKING PAPER SERIES
The impact of tax structure on investment: an empirical assessment for OECD countries
José Alves
REM Working Paper 058-2018
November 2018
REM – Research in Economics and Mathematics Rua Miguel Lúpi 20,
1249-078 Lisboa, Portugal
ISSN 2184-108X
Any opinions expressed are those of the authors and not those of REM. Short, up to two paragraphs can be cited provided that full credit is given to the authors.
The impact of tax structure on investment: anempirical assessment for OECD countries*
Jose Alves�
2018
Abstract
In the present empirical analysis we try to assess the impact of taxation on in-vestment growth. In particular, and by using gross fixed capital formation as a proxyfor investment, we intend to evaluate the impact of the taxation structure in invest-ment dynamics, in a short and a long-run perspectives. This empirical exercise wasconducted for all OECD countries, during the 1980-2015 period. Through panel dataeconometric techniques, we find optimal tax-investment threshold values, speciallyhigher for short-term than for long-term evolution. Also, we find optimal incometaxation rounding 9%, in percentage of GDP, an average optimal value 12.7% forconsumption taxes to promote annual investment growth.
Keywords: Investment Growth; Tax systems; Fiscal Policy; Optimal taxationJEL: D25; E62; H21; O47
*The opinions expressed herein are those of the author and do not necessarily reflect those of hisemployers. Any remaining errors are the author’s sole responsibility.
�ISEG/UL - Universidade de Lisboa, Department of Economics; REM – Research in Eco-nomics and Mathematics, UECE – Research Unit on Complexity and Economics. UECE issupported by FCT (Fundacao para a Ciencia e a Tecnologia, Portugal) Corresponding author:[email protected].
1
1 Introduction
It is trivial among academics that investment is crucial to promote long-run economicgrowth. Those decisions enable to sustain consumption in the long-run by applying theeconomic productive factors in both old and new economic production processes. Therefore,taxation can jeopardize the investment decisions. In particular, the raise of revenues ofboth income and consumption of individual and the taxation on firms can, on one hand,reduce the level of aggregate consumption, and, on another hand, decrease the investmentprofitability rates, through the reduction of aggregate demand that those investments areexpecting to face.
However, when the investment levels are beyond those, which promote the optimal con-sumption balanced path, it is imperative to promote the reduction of investment decisions.This happens when the condition of economic dynamic efficiency is not verified, i.e., whenthe return rate on capital exceeds the economic growth rate. More specifically, it is veri-fied a non-optimal level of investment when the marginal product of capital is lesser thanthe economic growth rate, as illustrated in several economic exogenous growth theories as,for e.g., in Solow (1956)-Swan (1956) and Ramsey (1928)-Cass (1965)-Koopmans (1963).In contrast to this perspective, when investment levels are below the optimal levels thatguarantee a sustainable growth path, there are some viewpoints that claim for governmentintervention, through the public spending and investment increasing. There are empir-ical evidences that sustain the possibility of the raise of public investment leading to acrowding-in effect in private investment and, therefore, increases in the aggregate invest-ment levels (Afonso and St. Aubyn (2009), for e.g.). Therefore, it is essential to analysethe taxation effects on investment dynamics, namely, what tax items can jeopardize grosscapital formation decisions and those that can promote it, in a short and long run per-spectives. Yet, in this study, it is our intention to compute possible non-linear relationsbetween tax items and gross fixed capital formation. In addition, we always consider theoverall government expenditures and that the revenues collected by taxes are reintroducedin the economic circuit.
This study is organised in the following sections: section 2 provides a brief reviewon the existing literature regarding the causalities of taxation on investment; section 2.1highlights the applied methodology, and the databases used in this analysis; section 3details the obtained results, and, lastly, section 4 summarizes our conclusions.
2 Literature Review
Some papers address the impact of tax policies on investment behaviour, namely thecorporate income tax and its effects on investment decision-making process. We can high-light, for instance, the study conducted by Da Rin et al. (2010) which makes use of paneldata techniques to assess the impact of taxation on firms for a set of more than 2.5 millionfirms in 17 European countries, and for the period between 1997 and 2004. The authorsconclude that a corporate tax reduction is related with a decreasing capital-labour ratio,
1
more specifically, the impact of corporate taxes is stronger on capital than labour firm’sproportion. Yet, as the authors highlight, a tax reduction is desired to promote the entranceof firms on the market; however, this policy can also favour the entrance of less financialrobust firms. The same conclusion regarding the effect of corporate taxation and the mar-ket entry is reached in Braunerhjelm and Eklund (2014). In fact, the authors verified thata 10% reduction in corporate taxation increases the market entry in 3%. Complementingthe previous conclusions, Rin et al. (2011) concludes for a non-linear relationship betweentax and firms entrance on the market.
For 14 developed countries in the 1982-2007 period, Bond and Xing (2015) find anegative relationship between taxes on firms and its effects on firm’s capital–output ratios.The authors develop an econometric specification for both short-run, through the availabledata, and long-run, derived from a constant elasticity of substitution neoclassical model ofinvestment, and find negative impacts of a 1% increase in firm’s taxation on capital-outputratios between -0.3% and -0.7%. These results are also corroborated by Djankov et al.(2010) for a sample of 85 countries in 2004. Moreover, it is also find that manufacturingactivities are more exposed to detrimental effects of corporate taxation. Another studythat corroborates the previous conclusions is the Mukherjee et al. (2017). Besides the factof finding a negative impact of taxes on corporate income and R&D activities, the authorsalso conclude that the consequence of higher taxes is a lesser supply of new goods andservices into the market economy. On the other hand, Galindo and Pombo (2011) find thatcorporate taxes affect more big firms than small and medium size firms in what respects toinvestment decisions and productivity. In addition, Brandstetter and Jacob (2013) applya difference-in-differences approach to assess the corporate tax on investment dynamicsfor the German case, and find heterogeneous responses, i.e., a cut in corporate tax couldlead to investment growth for domestic owned firms in a higher extent when comparedwith foreign-owned corporations. Yet, Baliamoune-Lutz and Garello (2014) found that,contrarily to high-income individuals, tax progressivity stimulates the entrepreneur’s entryfor the low-to-average income individuals.
In what respects to productivity, for a set of 11 European countries between 1996and 2005, Gemmell et al. (2016) conclude that while higher statutory corporate tax ratesimpact negatively in productivity levels of small firms, while productivity of bigger firmsis only affected by effective marginal tax rates. Additionally, Langenmayr et al. (2015)shows the existence of optimal corporation tax structure depending on the competitiondegree. In fact, the authors conclude that when the market competition degree is low,higher taxes favour high productivity firms; when the competition degree tends to thecompetitive market conditions and firms’ profit taxes are low, the low-productivity firmstend to be favoured.
By analysing the effects of consumption taxes on corporate investment decisions, Jacobet al. (2017) conclude that this source of taxation is also detrimental for firm’s investmentpurposes. The results reached by the authors led to the conclusion that the detrimentaleffect of consumption taxation is stronger for firms with a higher demand elasticity de-gree, besides of having a higher exposure to the domestic final consumer’s degree and tofinancial restrictions. Another topic is the tax burden and its relationship with risk-taking
2
decisions for firms’ investment. Regarding this, Ljungqvist et al. (2017) conclude thatthe response to a tax change is not symmetric. In fact, the results suggest that a taxincrease is accompanied by a R&D activities’ reduction, among others activities. The au-thors also conclude that only low financial leverage firms react to tax cuts, in what respectrisk increasing investment decisions. In addition, a study conducted by Ljungqvist andSmolyansky (2016) about the effect of corporate taxation on employment and income, inthe United States between 1970 and 2010, concluded that while corporate taxes’ reductionhas little impact on economic growth, tax cuts during an economic contraction can raiseboth levels of employment and income.
2.1 Methodology and Data
In this study, we consider that the investment growth, through the growth of grossfixed capital formation, is a function of taxation composition, denoted by T , of the type∆I = F (T ).
∆Ii,t = αi,t+β0,i,tgyi,t +∑
β1,i,tτt+βjxi,t+νi+ηt+εi,t, j = 1, 2, t = 1, ..., T, i = 1, ..., N (1)
where ∆Ii,t is the investment growth rate (annual or 5-years average) - in our case we usegross fixed capital formation growth rate as a proxy variable to measure investment growth-, gyi,t−1
is the real per capita GDP growth rate, τt represents each tax item, as a share ofGDP, xi,t is an independent variable belonging to the control variables’ set, νi and ηt are,respectively, the country and time-specific effects, and εi,t is the error term of the whitenoise-type.
Additionally, we introduce a squared term for each tax component to evaluate theexistence of non-linearity effects of tax structure on investment decisions, as demonstratedin equation (2).
∆Ii,t = αi,t+β0,i,tyi,t−1+∑
β1,i,tτt+∑
β2,i,tτ2t +βjxi,t+νi+ηt+εi,t, t = 1, ..., T, i = 1, ..., N
(2)By deriving equation 2 we obtain equation 3:
∂∆Ii,t∂(τi,t, τ 2
i,t)=∂(αi,t + β0,i,tyi,t−1 +
∑β1,i,tτt +
∑β2,i,tτ
2t + βixi,t + νi + ηt + εi,t)
∂(τi,t, τ 2i,t)
(3)
Each tax threshold is computed by equalizing equation (3) to zero, as in equation (4):
0 = β1 + 2β2,i,tτt ⇔ τt =−β1,i,t
2β2,i,t
(4)
Therefore, if we obtain a significant negative signal for β2,i,t we have a concave relation-
3
ship between a tax item and the investment dynamic, translating into an optimal valuefor taxation to maximize investment. On the other hand, a convex relationship througha positive coefficient for β2,i,t translates into a value that hampers investment growth de-cisions. Therefore, in the empirical results section when we get convex relations we willhighlight that coefficient to differentiate between maximum and minimum optimal levels.
The model herein computed considers the period between 1980 and 2015, for the overallOECD countries: Australia (AUS), Austria (AUT), Belgium (BEL), Canada (CAN), Chile(CHL), Czech Republic (CZE), Denmark (DNK), Estonia (EST), Finland (FIN), France(FRA), Germany (DEU), Greece (GRC), Hungary (HUN), Iceland (ISL), Ireland (IRL),Israel (ISR), Italy (ITA), Japan (JPN), South Korea (KOR), Latvia (LVA), Luxembourg(LUX), Mexico (MEX), the Netherlands (NLD), New Zealand (NZL), Norway (NOR),Poland (POL), Portugal (PRT), Slovak Republic (SVK), Slovenia (SVN), Spain (ESP),Sweden (SWE), Switzerland (CHE), Turkey (TUR), United Kingdom (GBR) and UnitedStates (USA).
The final database used in our analysis includes data from several sources: the PPP percapita GDP (realgdppc), the public debt (debt) and the total government spending (tot-exp), both in ratio of GDP, and output gap, in percent of potential GDP (outputgap) arefrom World Economic Outlook (IMF); taxes on income, profits and capital gains of indi-viduals (taxinc), taxes on income, profits and capital gains of corporates (taxfirms), socialsecurity contributions (ssc), taxes on payroll and workforce (taxpayroll), taxes on property(taxprop), taxes on goods and services (taxvat), gross fixed capital formation (gfcf ) andits growth rate (gfcfgr) were retrieved from OECD.Stats database, the age dependencyratio, as percentage of active population (ageratio), deposit interest rate (depositrate), netforeign direct investment-to-GDP ratio (foreigninvestment), and the GDP percentage ofhousehold final consumption expenditure (hconsggdp) are from the World Development In-dicators (WDI). Population in millions (pop) and the real total factor productivity (rtfpna)are from Feenstra et al. (2015) data. Lastly, while liquid liabilities-to-GDP ratio (llgdp) isbased on International Financial Statistics (IFS), IMF. The table 1 presents the summarystatistics for each variable used in our regressions1.
For the coefficients estimation, we resort to panel data techniques, throughout theapplication of OLS, OLS-Fixed Effects (FE), Generalized Method of Moments (GMM) andRobust Least Squares (RLS) methodologies. With the exception of RLS, those estimationsassume the white diagonal covariance matrix hypothesis. Additionally, we estimate bothequations (1) and (2) for both annual and 5-years average growth rates. Lastly, we will onlydiscuss the threshold existence when the coefficients of each tax items present statisticalsignificance for both linear and square term items tax regressors, for a minimum of 90%confidence interval.
Table 1: Summary statistics of the variables set for investment regressions, 1980-2015.
realgdppc taxinc taxfirms ssc taxpayroll taxprop taxvat
Mean 24.448 8.820 2.806 8.345 0.369 1.745 10.588
1For reasons of parsimony, the results of realgdppc are expressed in thousands of USD.
4
Std dev 14.313 4.635 1.500 4.981 0.728 1.003 3.046Max 101.054 26.780 12.594 19.173 5.661 7.334 18.730Min 2.184 0.873 0.261 0.000 0.000 0.074 2.979Obs. 1195 1106 1106 1137 1137 1137 1137
gfcf gfcfgr depositrate ageratio debt foreigninvestment rtfpna
Mean 23.161 3.314 9.253 51.287 55.728 3.645 0.941Std dev 4.091 8.917 25.364 6.931 35.901 10.487 0.123Max 39.404 45.119 682.53 96.457 242.113 252.308 1.539Min 11.546 -47.761 -0.180 36.323 3.664 -58.323 0.472Obs. 1174 1164 1055 1260 943 1120 1173
totexp pop hconsggdp outputgap
Mean 42.621 33.531 56.382 -0.319Std dev 9.657 52.235 7.069 2.85Max 68.436 319.449 79.551 14.911Min 14.244 0.228 29.918 -11.437Obs. 977 1173 1174 851
3 Results
3.1 Short-run effects of taxation on investment dynamics
The short-run analysis for equation (1), i.e, without the tax items square terms, evi-dence a negative relation between all type of taxes and investment increasing, as it wasexpected, with the exception of payroll taxes, which appears to have a positive correlationbetween the revenue levied by this type of taxes and the investment decisions. In detail,while by an increment of 1% of revenues collected from payroll taxes, in proportion of GDP,the gross capital formation increases by 2%, the negative impact of a one unit increasingin revenues by all mentioned taxes vary approximately between 0.09% and 0.66%.
Moreover, we verify a negative impact in the following variables: deposit interest rates(by about 0.4%), household consumption (0.2%). On contrary, we verify the positive im-pact of foreign investment, always greater than 0.1%. Looking in detail at the governmentdebt growth impact, it seems that this variable crowds out aggregate investment. Onthe other hand, and while public expenditures growth and age dependency ratio do notevidence to influence investment decisions, output gap reveals to be favourable to grossfixed capital formation, between 0.27% and 0.74%. This positive effect can be explainedby the fact that when the economy is overheating, inflation pressures could decrease theamortization time of the investment, and, consequently, increase its profitability.
In what respects to tax items thresholds for investment decision-making, through thegrowth rate of gross fixed capital formation, it is possible to verify the non-existence ofthis thresholds for taxes on payroll and taxes on property. Regarding the other tax items,we reach to a value of 10.65% for taxes on individual income on GDP, which translatesthe maximum value that promotes investment, while we assist to minimizing values forsocial security contributions of 12.09%, and 7.37% and 14.18%, on average, for tax onfirms and on consumption, respectively. The values reached for the last three tax itemsevidence their minimum revenue, in GDP proportion, to promote investment growth. The
5
above-mentioned results are presented in table 2.
Table 2: Linear and non-linear short-run impact results of taxation structure on investmentdecisions.
OLS OLS-FE GMM RLS
(1) (2) (3) (4) (5) (6) (7) (8)
OLS OLS-FE GMM RLS
(1) (2) (3) (4) (5) (6) (7) (8)
∆realgdppc 0.004*** 0.003*** 0.003*** 0.003*** 0.004*** 0.005*** 0.004*** 0.004***(0.000) (0.000) (0.001) (0.001) (0.001) (0.001) (0.000) (0.000)
taxinc−1 -0.144** -0.069 0.131 1.491** -0.127 0.037 -0.128** -0.182(0.070) (0.215) (0.199) (0.577) (0.112) (0.367) (0.061) (0.175)
taxinc2−1 -0.007 -0.070** -0.015 -0.001(0.009) (0.031) (0.017) (0.008)
taxfirms−1 -0.442** -1.580*** 0.277 -0.595 -0.561*** -3.859*** -0.289** -1.032**(0.176) (0.568) (0.272) (0.779) (0.263) (1.488) (0.140) (0.427)
taxfirms2−1 0.107** 0.070 0.279** 0.066*(0.044) (0.060) (0.115) (0.039)
ssc−1 -0.121** -0.436** -0.047 -2.007** -0.151 -0.703 -0.086* -0.276(0.058) (0.209) (0.230) (1.013) (0.095) (0.437) (0.050) (0.180)
ssc2−1 0.017 0.083** 0.028 0.010(0.011) (0.039) (0.021) (0.010)
taxpayroll−1 -0.324 0.420 1.990** 3.002 -0.065 -2.013 -0.284 0.107(0.262) (0.929) (0.968) (1.904) (0.381) (1.546) (0.277) (0.934)
taxpayroll2−1 -0.365 -0.304 0.729 -0.144(0.356) (0.540) (0.682) (0.383)
taxprop−1 -0.571* -0.588 -0.211 0.775 -0.036 3.822 -0.013 1.139(0.295) (1.025) (0.508) (1.521) (0.649) (2.793) (0.255) (0.770)
taxprop2−1 0.042 -0.118 -0.658 -0.208(0.166) (0.214) (0.443) (0.141)
taxvat−1 -0.612*** -2.640*** 0.413 0.291 -0.103 -2.691* -0.663*** -1.982***(0.163) (0.635) (0.363) (1.127) (0.3) (1.519) (0.133) (0.538)
taxvat2−1 0.099*** 0.016 0.111 0.066**(0.033) (0.057) (0.072) (0.026)
gfcf−1 -0.521*** -0.615*** -1.283*** -1.337*** -0.077 -0.340 -0.505*** -0.599***(0.104) (0.123) (0.148) (0.159) (0.111) (0.208) (0.072) (0.079)
gfcfgr−1 0.242*** 0.242*** 0.225*** 0.221*** 0.429** 0.267 0.259*** 0.257***(0.057) (0.058) (0.061) (0.061) (0.183) (0.169) (0.032) (0.032)
∆depositrate -0.426* -0.435* -0.436* -0.414* 1.063 -1.323 -0.583*** -0.606***(0.241) (0.238) (0.256) (0.251) (1.233) (1.526) (0.160) (0.159)
ageratio−1 -0.016 -0.053 -0.065 -0.029 -0.044 -0.029 -0.027 -0.058(0.067) (0.071) (0.137) (0.144) (0.116) (0.104) (0.053) (0.055)
∆debt -0.275*** -0.303*** -0.189** -0.200** -0.487* -0.543** -0.264*** -0.295***(0.078) (0.082) (0.082) (0.084) (0.258) (0.218) (0.047) (0.047)
rtfpna -12.705*** -12.809*** -6.562 -2.947 2.829 -0.873 -10.328*** -10.345***(3.769) (3.884) (5.536) (6.308) (4.246) (5.567) (2.677) (2.770)
∆totexp−1 -0.148 -0.160 -0.088 -0.081 0.315 -0.139 0.053 0.032(0.159) (0.163) (0.154) (0.157) (0.654) (0.602) (0.104) (0.104)
log(pop) -0.497 -0.580 61.785*** 63.702*** 0.228 -0.501 -0.926*** -0.997***(0.397) (0.396) (10.841) (11.771) (0.878) (0.974) (0.237) (0.243)
foreigninvestment−1 -0.075*** -0.066*** -0.070*** -0.066*** 0.028 0.009 -0.075*** -0.073***(0.017) (0.016) (0.017) (0.017) (0.087) (0.091) (0.015) (0.015)
hconsumption−1 0.039 0.013 -0.214** -0.242** 0.07 0.000 0.036 0.024(0.027) (0.028) (0.093) (0.094) (0.048) (0.049) (0.027) (0.028)
outputgap 0.369** 0.466*** 0.690*** 0.736*** -0.849** 0.121 0.271*** 0.352***(0.157) (0.157) (0.159) (0.170) (0.395) (0.565) (0.091) (0.093)
Tax thresholdstaxinc - - - 10.65% - - - -
6
taxfirms - 7.38% - - - 6.92% - 7.82%ssc - - - 12.09% - - - -taxpayroll - - - - - - - -taxprop - - - - - - - -taxvat - 13.33% - - - - - 15.02%
R-squared 0.540 0.554 0.680 0.687 0.343 0.477 0.401 0.414DW-Stat 1.856 1.875 2.000 2.018 2.110 1.949 n.a. n.a.Obs. 529 529 529 529 473 473 529 529
Notes: *, ** and *** represent statistical significance at levels of 10%, 5% and 1% respectively. The robust standard errors arein brackets. The White diagonal covariance matrix is used in order to assume residual heterokedasticity, with the exceptionfor RLS technique. The DW-statistic is the Durbin-Watson statistic. The non-bold and bold values express, respectively,maximum and minimum optimal tax items levels.
3.2 Long-run effects of taxation on investment dynamics
In a long-run perspective, and in a linear relationship (see regressions (9), (11), (13) and(15)), the results obtained highlight similar patterns for the verified short-run effects ontax items and investment growth, with the exception for the tax on property that appearsto be not relevant to determine investment decisions. In addition, we can observe that thevalues presented in tables 2 and 3 highlight similar magnitudes for the tax items coefficientsin investment growth.
Regarding the other control variables, and similar to the short-run perspective, popu-lation size evidences a contradictory signal. Moreover, household consumption and outputgap variables highlight a non-clear effect on investment dynamics, since these last twovariables also present different signals, depending on the econometric technique used.
Yet, it seems that the deposit interest rates had a negative effect on investment ina long-term perspective. Specifically, by a 1 p.p. increase in deposit interest rates, theinvestment tend to decrease between 0.4 p.p. and 0.6 p.p., approximately. Additionally, thegovernment spending variation also seems to gain importance in the long-term, presentinga slightly negative impact on capital formation.
In what respects to the analysis of the non-linear relationships of tax items on invest-ment decision, and by computing the consequent existing tax items-to-investment thresh-olds, we find maximum values of 6.27% and 9.19% for taxation on firms’ profits and forconsumption taxes, respectively. On the opposite side, we find a minimum threshold valuefor social security contributions of 11.35%. In the long run, we can also observe that theoptimal for taxes on payroll disappears. The above-discussed results may be observed indetail in table 3.
Table 3: Linear and non-linear long-run impact results of taxation structure on investmentdecisions.
OLS OLS-FE GMM RLS
(9) (10) (11) (12) (13) (14) (15) (16)
∆realgdppc 0.000** 0.000** 0.003*** 0.001** 0.004*** 0.001** 0.004*** 0.000*(0.000) (0.000) (0.001) (0.000) (0.001) (0.000) (0.000) (0.000)
taxinc−1 -0.042 -0.085 0.131 0.740 -0.127 -0.263 -0.128** 0.021
7
(0.051) (0.091) (0.199) (0.571) (0.112) (0.169) (0.061) (0.114)taxinc2−1 0.003 -0.042 0.012* -0.004
(0.004) (0.029) (0.007) (0.005)taxfirms−1 -0.098 -0.197 0.277 0.489 -0.561** -0.601 -0.289** -0.828***
(0.107) (0.458) (0.272) (0.747) (0.263) (0.734) (0.140) (0.278)taxfirms2−1 0.014 -0.028 0.041 0.066***
(0.034) (0.053) (0.055) (0.025)ssc−1 -0.069** -0.227** -0.047 -0.417 -0.151 -0.150 -0.086* -0.255**
(0.034) (0.091) (0.230) (0.662) (0.095) (0.172) (0.050) (0.117)ssc2−1 0.010* 0.008 0.006 0.009
(0.005) (0.023) (0.010) (0.006)taxpayroll−1 -0.168 0.170 1.990** 1.156 -0.065 0.221 -0.284 -0.164
(0.160) (0.490) (0.968) (1.352) (0.381) (0.986) (0.277) (0.608)taxpayroll2−1 -0.215 -0.467 -0.334 -0.044
(0.185) (0.503) (0.415) (0.249)taxprop−1 -0.258 -0.971 -0.211 -0.816 -0.036 -1.381 -0.013 0.447
(0.181) (0.868) (0.508) (1.838) (0.649) (1.260) (0.255) (0.501)taxprop2−1 0.137 0.122 0.190 -0.052
(0.124) (0.219) (0.198) (0.092)taxvat−1 -0.271** -0.183 0.413 1.985*** -0.103 0.298 -0.663*** -0.499
(0.109) (0.397) (0.363) (0.571) (0.300) (0.853) (0.133) (0.350)taxvat2−1 -0.007 -0.108*** -0.040 0.022
(0.021) (0.029) (0.035) (0.017)gfcf−1 -0.114* -0.100 -1.283*** -0.259*** -0.077 0.096 -0.505*** -0.083
(0.062) (0.069) (0.148) (0.091) (0.111) (0.115) (0.072) (0.052)gfcfgr−1 0.274*** 0.267*** 0.225*** 0.214*** 0.429** 0.265*** 0.259*** 0.292***
(0.034) (0.036) (0.061) (0.036) (0.183) (0.100) (0.032) (0.021)∆Depositrate -0.648*** -0.644*** -0.436* -0.468*** 1.063 -0.388 -0.583*** -0.573***
(0.124) (0.191) (0.256) (0.131) (1.233) (0.815) (0.160) (0.104)ageratio−1 -0.064 -0.053 -0.065 -0.052 -0.044 0.089** -0.027 -0.071**
(0.040) (0.032) (0.137) (0.112) (0.116) (0.045) (0.053) (0.036)∆debt -0.239*** -0.238*** -0.189** -0.150*** -0.487* -0.631*** -0.264*** -0.245***
(0.050) (0.057) (0.082) (0.041) (0.258) (0.124) (0.047) (0.031)rtfpna -4.849** -4.240* -6.562 0.082 2.829 1.089 -10.328*** -4.070**
(2.074) (2.551) (5.536) (4.644) (4.246) (3.229) (2.677) (1.803)∆totexp−1 -0.046 -0.050 -0.088 0.002 0.315 0.517 0.053 -0.125*
(0.091) (0.077) (0.154) (0.079) (0.654) (0.327) (0.104) (0.068)log(pop) -0.436** -0.421 61.785*** 14.650** 0.228 -0.203 -0.926*** -0.415***
(0.189) (0.299) (10.841) (7.225) (0.878) (0.391) (0.237) (0.158)foreigninvestment−1 -0.004 -0.002 -0.070*** -0.006 0.028 0.044 -0.075*** -0.003
(0.007) (0.006) (0.017) (0.009) (0.087) (0.047) (0.015) (0.010)hconsumption−1 0.056*** 0.053*** -0.214** -0.093 0.070 0.057** 0.036 0.026
(0.018) (0.018) (0.093) (0.102) (0.048) (0.028) (0.027) (0.018)outputgap -0.144* -0.138 0.690*** -0.085 -0.849** -0.433 0.271*** -0.214***
(0.083) (0.103) (0.159) (0.091) (0.395) (0.312) (0.091) (0.061)
Tax thresholdstaxinc - - - - - - - -taxfirms - - - - - - - 6.27%ssc - 11.35% - - - - - -taxpayroll - - - - - - - -taxprop - - - - - - - -taxvat - - - 9.19% - - - -
R-squared 0.476 0.481 0.717 0.730 0.171 0.180 0.360 0.366DW-Stat 1.124 1.127 1.034 1.095 1.692 1.665 n.a. n.a.Obs. 529 529 529 529 473 473 529 529
Notes: *, ** and *** represent statistical significance at levels of 10%, 5% and 1% respectively. The robust standard errors arein brackets. The White diagonal covariance matrix is used in order to assume residual heterokedasticity, with the exceptionfor RLS technique. The DW-statistic is the Durbin-Watson statistic. The non-bold and bold values express, respectively,maximum and minimum optimal tax items levels.
8
4 Concluding Remarks
In this study, we developed an empirical model to assess possible relations betweentax composition, in GDP proportion, and investment growth. To perform this exercise weused gross fixed capital formation as a proxy variable for investment, and our empiricalanalysis resorted to panel data techniques to analyse tax effects in both short and long-term. In addition, we assessed the existence of non-linear relationships between tax itemsand investment growth.
The obtained results evidenced the existence of some tax-to-GDP thresholds. Namely,there were some optimal values of tax items, in share of GDP, which maximizes the in-vestment decisions. In particular, and regarding the short-run, we reach a maximizingthreshold, which promotes long-run investment growth of 10.65% for taxes on individualincome. On the other hand, in the end, we conclude for maximum threshold levels for taxeson firms’ profits and taxes on consumption of goods and services of 6.27% and 9.19%, re-spectively, to promote investment growth. Moreover, we found a minimum threshold of11.35% of social security contribution.
In what respects the short-run, we found only one maximum threshold of 10.65% fortaxes on individual income, while it is found a minimum threshold value of 12.09% for socialcontributions. In addition, we found minimum threshold values, on average, of 7.37% and14.18% for profits and consumption of goods and services taxes, respectively.
Lastly, and resorting to the results presented in table 1 which allow the comparison ofthe short and long-term results for each tax items from the econometric regressions, wewere also able to conclude that there are some fiscal space to raise some taxes to promotehigher investment growth rates, mainly in a short-run perspective. In particular, with theexception of payroll taxes and taxes on property on firms, we did not find any optimalvalue - it seems that the raise of the other taxes tends to be a favourable political measureto promote investment growth. Moreover, if we sum all the optimal threshold tax itemsvalues with the historical average recorded for taxes, we conclude that taxation over theGDP should be around 46.41% and 37.75% in the short-run and long-run, respectively.The table 4 summarizes our main findings regarding average tax threshold values.
9
Table 4: Summary of tax items threshold values for investment decisions.
Short-run Long-run Mean
taxinc 10.65% - 8.82%
taxfirms 7.37% 6.27% 2.81%
ssc 12.09% 11.35% 8.35%
taxpayroll - - 0.37%
taxprop - - 1.75%
taxvat 14.18% 9.19% 10.59%
Notes: The non-bold and bold values, presented in the short-run and long-run columns express maximum and minimumoptimum levels, respectively. The values expressed in italics represent average values.
References
Afonso, A. and St. Aubyn, M. (2009). Macroeconomic Rates of Return of Public andPrivate Investment: Crowding-in and Crowding-out Effects. The Manchester School,77:21–39.
Baliamoune-Lutz, M. and Garello, P. (2014). Tax structure and entrepreneurship. SmallBusiness Economics, 42(1):165–190.
Bond, S. and Xing, J. (2015). Corporate taxation and capital accumulation: Evidencefrom sectoral panel data for 14 OECD countries. Journal of Public Economics, 130(Sup-plement C):15 – 31.
Brandstetter, L. and Jacob, M. (2013). Do corporate tax cuts increase investments? Arqus-Diskussionsbeitrage zur quantitativen Steuerlehre 153.
Braunerhjelm, P. and Eklund, J. E. (2014). Taxes, tax administrative burdens and newfirm formation. Kyklos, 67(1):1–11.
Cass, D. (1965). Optimum Growth in an Aggregative Model of Capital Accumulation. TheReview of Economic Studies, 32(3):233–240.
Da Rin, M., Sembenelli, A., and Di Giacomo, M. (2010). CORPORATE TAXATION ANDTHE SIZE OF NEW FIRMS: EVIDENCE FROM EUROPE. Journal of the EuropeanEconomic Association, 8(2-3):606–616.
Djankov, S., Ganser, T., McLiesh, C., Ramalho, R., and Shleifer, A. (2010). The Effectof Corporate Taxes on Investment and Entrepreneurship. American Economic Journal:Macroeconomics, 2(3):31–64.
10
Feenstra, R. C., Inklaar, R., and Timmer, M. P. (2015). The Next Generation of the PennWorld Table. American Economic Review, 105(10):3150–3182.
Galindo, A. J. and Pombo, C. (2011). Corporate taxation, investment and productivity:A firm level estimation. Journal of Accounting and Taxation, 5(7):158–161.
Gemmell, N., Kneller, R., McGowan, D., Sanz, I., and Sanz-Sanz, J. F. (2016). CorporateTaxation and Productivity Catch-Up: Evidence from European Firms. The ScandinavianJournal of Economics, pages 0–34.
Jacob, M., Michaely, R., and Muller, M. A. (2017). Consumption taxesand corporate investment. Technical report. Available at SSRN:https://ssrn.com/abstract=2800146https://ssrn.com/abstract=2800146.
Koopmans, T. C. (1963). On the Concept of Optimal Economic Growth. Technical Report163, Cowles Foundation for Research in Economics, Yale University.
Langenmayr, D., Haufler, A., and Bauer, C. J. (2015). Should tax policy favor high- orlow-productivity firms? European Economic Review, 73(Supplement C):18 – 34.
Ljungqvist, A. and Smolyansky, M. (2016). To Cut or Not to Cut? On the Impact ofCorporate Taxes on Employment and Income. Finance and Economics Discussion Series2016-6, Board of Governors of the Federal Reserve System (U.S.).
Ljungqvist, A., Zhang, L., and Zuo, L. (2017). Sharing Risk with the Government: HowTaxes Affect Corporate Risk Taking. Journal of Accounting Research, 55(3):669–707.
Mukherjee, A., Singh, M., and Zaldokas, A. (2017). Do corporate taxes hinder innovation?Journal of Financial Economics, 124(1):195 – 221.
Ramsey, F. P. (1928). A Mathematical Theory of Saving. The Economic Journal,38(152):543–559.
Rin, M. D., Giacomo, M. D., and Sembenelli, A. (2011). Entrepreneurship, firm entry, andthe taxation of corporate income: Evidence from Europe. Journal of Public Economics,95(9):1048 – 1066.
Solow, R. M. (1956). A Contribution to the Theory of Economic Growth. The QuarterlyJournal of Economics, 70(1):65–94.
Swan, T. W. (1956). Economic Growth and Capital Accumulation. Economic Record,32(2):334–361.
11
